Improving methods for genotypic drug resistance testing in
Mycobacterium tuberculosis
Zandile Cleopatra Mlamla
Thesis presented in partial fulfilment of the requirements for the degree Master of Science in Medical Sciences (Medical Biochemistry) at the University
of Stellenbosch
Supervisor: Prof Thomas C.Victor Co-supervisors: Prof Robin M. Warren
Dr Gail E. Louw Faculty of Health Sciences
Division of Molecular Biology and Human Genetics, Department of Biomedical Sciences
ii
Declaration
I, Zandile Cleopatra Mlamla, hereby declare that the work contained in this thesis is my own original work and that I have not previously in its entirety or in part submitted it at any university for a degree
Signature: ... Date: 1 December 2010
"Without Him, I am nothing and can do nothing of value. He is my only Source."
Copyright © 2011 Stellenbosch University All rights reserved
iii
Summary
An important next step to Tuberculosis control relies on the translation of basic science and modern diagnostic techniques into primary health care clinics. These assays must be rapid, inexpensive, interpretation of results must be easy and they must be simple so that a healthcare worker with limited training can perform the tests under safe conditions. This study consists of four aims. The first aim was to develop a methodology to sterilize sputum specimens for rapid TB diagnosis and drug resistance testing. Candidate bactericides were identified from the literature, and tested for their bactericidal activity in Mycobacterium tuberculosis. We identified ultraseptin®aktiv as a powerful bactericidal agent which sterilizes sputum specimens for subsequent safe handling prior to light emitting diode microscopy and it also provides a DNA template for PCR-based tests. An algorithm has been proposed for the processing of specimens and rapid diagnosis of TB and drug resistant TB while patients wait for results.
Recently, the World Health Organization has endorsed the MTBDRplus test for diagnosis of TB and drug resistant TB. However genotypic tests may have more problems than anticipated. With the HIV pandemic, an increase of non-tuberculous mycobacteria has been reported. The sensitivity of genotypic tests in specimens with underlying non-tuberculous mycobacterial species therefore requires further evaluation. This study therefore also aimed at determining the reliability of the MTBDRplus assay for detection of drug resistant TB where non-tuberculous bacterial load is high. Clinically relevant non-tuberculous mycobacterium DNA and DNA from a multi-drug resistant TB isolate were obtained. Ratios of the different NTM with the MDR-TB DNA were made and subjected to the MTBDRplus assay. Known mix NTM and TB infected clinical isolates and sputum sediments were also evaluated for TB and drug resistance detection on the MTBDRplus assay. Under these conditions, this study provides evidence that the MTBDRplus test cannot reliably detect TB and drug resistance TB in specimens with underlying non-tuberculous mycobacteria.
iv Thirdly, to evaluate the sensitivity of the MTBDRplus assay for detecting drug resistance in hetero-resistant isolates, ratios were made using purified DNA from an MDR and pan-susceptible TB isolate. The MTBDRplus assay was then performed on the different ratios. We report that the MTBDRplus assay can efficiently detect wild type DNA in genes associated with resistance during the early evolution of drug resistance. However, in the later stage during treatment when both the wild type and mutants are present, the detection limit for the mutant DNA was 1:55. Due to these results, the MTBDRplus assay should still be further improved or other tests should be developed to address these limitations.
And finally to combat cross amplicon contamination during the final steps of genotypic detection with the MTBDRplus assay, a proof of concept for a patentable closed tube line probe device was proposed on the 4th aim. This device can be improved to enable automated drug resistance genotyping of multiple specimens.
The results of this study highlight the need for a sensitive inexpensive point of care drug resistance test that does not require intensive training.
v
Opsomming
'n Belangrike volgende stap om Tuberkulose te beheer is om basiese wetenskap resultate te gebruik sodat moderne diagnose tegnieke ontwikkel kan word wat in primêre gesondheidsorg klinieke toegepas kan word. Hierdie toetse moet vinnig, goedkoop, en die interpretasie van resultate moet maklik wees. Die toetse moet eenvoudig wees sodat 'n gesondheidswerker met beperkte opleiding die toetse onder veilige omstandighede kan uitvoer. Hierdie studie bestaan uit vier doelwitte, waarvan die eerste was om „n metode te ontwikkel vir die sterilisasie van sputum monsters vir vinnige TB diagnose en die toesting van middelweerstandigheid. Kandidaat kiemdodende middels was geïdentifiseer vanaf die literatuur en die middels se kiekdodende aktiviteit was getoets op Mycobacterium tuberculosis. Ons het ultraseptin®aktiv geïdentifiseer as 'n kragtige kiemdodende middel wat bakteria in sputum monsters steriliseer vir veilige hantering voordat diagnose met n lig uitstralende diode mikroskopie gedoen kan word. Hierdie behandeling met ultraseptin®aktiv bied ook 'n DNA templaat vir PCR-gebaseerde toetse. 'n Algoritme is voorgestel vir die hantering van monsters en die vinnige diagnose van sensitiewe- en middel weerstandige Tuberkulose terwyl die pasiënte by die kliniek wag vir die resultate.
Onlangs het die Wêreld Gesondheid Organisasie die genotipiese MTBDRplus toets vir die diagnose van Tuberkulose en middel-weerstandige Tuberkulose onderskryf. Hierdie toets word tans op groot skaal in Suid Afrika gebruik. Dit kan egter wees dat genotipiese toetse baie meer probleme kan he as wat aanvanklik verwag is. Die HIV pandemie gaan toenemend gepaard met n toename van nie-tuberkulose mycobacteria. Die sensitiwiteit van genotipiese toetse op monsters met onderliggende nie-tuberkulose mikobakteriese spesies vereis dus verdere evaluasie. Die doel van hierdie studie was ook om die betroubaarheid van die MTBDRplus-toets te bepaal vir die opsporing van middelweerstandige TB waar die nie-tuberkulose backteriële lading hoog is. DNA van kliniese relevante nie-tuberkulose mikobakteria en multi-middelweerstige TB isolate was bekom. Verskillende verdunnigs van die spesifieke NTM DNA te same met die van MDR-TB DNA is gemaak en onderwerp aan die MTBDRplus toets. Bekende gemengde NTM- en TB geïnfekteerde kliniese isolate en sputum sedimente was ook geevalueer vir die opsporing van TB en middel weerstandigheid met die MTBDRplus toets. Hierdie studie verskaf bewyse dat die
vi MTBDRplus toets nie betroubaar is met die diagnose van sensitiewe- en middel weerstandige Tuberkulose in monsters met onderliggende nie-tuberkulose mycobacteria nie.
Verskillende verdunnings van gesuiwerde DNA van MDR en pan-sensitiewe TB isolate is gemaak om die sensitiwiteit van die MTBDRplus toets vir die opsporing van middelweerstandigheid te bepaal. Die MDRDRplus toets is gebruik met hierdie verdunnings. Resultate in hierdie studie toon dat die MTBDRplus toets effektief is met die identifisering van wilde-tipe DNA (dit beteken middel sensitief) in gene wat geassosieer word met middel weerstandigheid gedurende die vroeë ontwikkeling van weerstandigheid. Hier teenoor toon die resultate dat in die later stadium tydens behandeling, wanneer beide die wilde-tipe (sensitief) en mutante DNA (weerstandig) teenwoordig is, is die opsporingslimiet vir die mutante DNA maar 1:55. As gevolg van hierdie resultate raai ons aan dat die MTBDRplus toets nog verder verbeter moet word of dat ander toetse ontwikkel moet word om hierdie beperkinge aan te spreek.
Amplikon kruiskontaminasie kan n groot impak hê op die betroubaarheid van enige genotipiese diagnostiese toets. Die finale stappe van MTBDRplus toets behels die gebruik van 'n oop sisteem sodat kontaminasie maklik kan plaasvind. In die 4de doewit 'n konsep vir 'n patenteerbare
geslotebuis toestel ontwikkel en die resultate het getoon dat kontaminasie suksesvol uitgeskakel kan word. Hierdie toestel kan verbeter na 'n outomatiese apparaat verbeter word sodat die module genotipering van verskeie monsters moontlik kan maak.
Die resultate van hierdie studie beklemtoon die noodsaaklikheid van 'n sensitiewe goedkoop “point of care” diagnostiese toets wat nie intensiewe opleiding benodig nie.
vii
Acknowledgements
I wish to extend my sincere gratitude to the following people, without whom the completion of this thesis would have not been possible:
My principal supervisor, Prof Thomas Victor; and my co-supervisors, Prof Rob Warren and Dr Gail Louw, I would like to thank them for their continued support, guidance, and encouragement throughout the course of this project.
Prof Nico Gey van Pittius, Marianne de Kock, Annemie Jordaan, Elizabeth Streicher and Monique Williams for their mentorship, guidance, provision of samples and training on
mycobacterial culture techniques, MTBDRplus assay as well as genotypic analysis of specimens. The staff at the Division of Medical Microbiology and Immunology, Tygerberg Hospital for the collection of sputum specimens.
My fellow colleagues and friends: Margaretha, Suereta, Rozanne, Melanie, Philippa, Leanie, and Prudy for their support, encouragement and empathy during the course of the MSc study and writting of the thesis.
The Medical Research Council, Stellenbosch University and the Department of Molecular Biology and Human Genetics for financial support.
To my family, for their encouragement, support and understanding during the course of this study.
viii
List of Abbreviations
°C : Degree Celsius
μl : microlitres
ADC : Albumin dextrose catalase
AM : Amikacin
bp : base pairs
BSA : Bovine serum albumin
CAP : Capreomycin
CIP : Ciprofloxacin
dH2O : Distilled water
DNA : Deoxyribonucleic acid
dNTP : Deoxyribonucleotide triphosphate
EDTA : Ethylenediaminetetraacetic acid
EMB : Ethambutol ETH : Ethionamide EtOH : Ethanol FQ : Fluoroquinolone g : Grams INH : Isoniazid KAN : Kanamycin
KCI : Potassium chloride
KNO3 : Potassium nitrate
LAM : Latin-American and Mediterranean
LCC : Low Copy Clade
LJ : Lowenstein-Jensen
M. abscessus : Mycobacterium abscessus
M. avium : Mycobacterium avium
M. bovis : Mycobacterium bovis
ix
M. fortuitum : Mycobacterium fortuitum
M. intracellulare Mycobacterium intracellulare
M. kansasii : Mycobacterium kansasii
M. peregrinum : Mycobacterium peregrinum
M. terrae : Mycobacterium terrae
M. tuberculosis : Mycobacterium tuberculosis
MDR : Multi Drug Resistant
MIC : Minimum Inhibitory Concentration
ml : millilitres
mM : millimetres
NALC : N-acetyl-L-cysteine
NaOCl : Sodium hypochlorite
NaOH : Sodium hydroxide
OADC : Oleic acid/albumin/dextrose/catalase
OFX : Ofloxacin
OPA : Ortho-phthalaldehyde
PBS : Phosphate buffer saline
PCR : Polymerase chain reaction
RIF : Rifampicin
rpm : revolutions per minute
RNA : Ribonucleic acid
rRNA : Ribosomal RNA
SA : South Africa
SDS : Sodium dodecyl sulphate
SNP : Single nucleotide polymorphism
STR : Streptomycin
TB : Tuberculosis
Tm : Melting temperature
Tris : Trishydroxymethylaminomethane
XDR : Extensive Drug Resistant
x
List of Figures
Figure 3-1 Schematic representation of the bactericide protocol. ... 56
Figure 3-2 Single tube line probe assay prototype ... 62
Figure 3-3 Material used to construct the single tube line probe prototype ... 62
Figure 4-1 BCG Pasteur cells exposed to bactericides. ... 70
Figure 4-2 PCR amplified products of the rpoB gene on agarose gel after bactericide treatment ... 71
Figure 4-3 Mycobacterial stains of specimens exposed to OPA and Ultraseptin®aktiv ... 74
Figure 4-4 Efficiency of PCR amplification after bactericide treatment. ... 75
Figure 4-5 MTBDRplus assay results ... 78
Figure 4-6 MTBDRplus assay results for the heteroresistant ratio mixtures ... 81
Figure 4-7 MTBDRplus prototype of a single tube device. ... 83
Figure 4-8 Single tube MTBDRplus assay results ... 83
Figure 4-9 MTBDRplus prototype of a single tube device. ... 84
Figure 4-10 Proposed patentable prototype MTBDRplus device ... 84
xi
List of Tables
Table 2-1 Critical concentrations for drug susceptibility testing of M. tuberculosis isolates in
different media. ... 23
Table 2-2 Performance of phenotypic DST methods ... 24
Table 2-3 Gene(s) associated with drug resistance in M. tuberculosis ... 25
Table 2-4 Molecular assays with their respective overall accuracies for detection of drug resistance ... 26
Table 2-5 Factors that negatively impact on drug resistance detection by molecular techniques 27 Table 3-1 Selected bactericides ... 49
Table 3-2 Bacterial load and quality of sputum specimens. ... 52
Table 3-3 Primer sets for gene amplification ... 55
Table 3-4 Selected non-tuberculous mycobacterial species and drug resistant isolates. ... 57
Table 3-5 Primer sets for gene amplification ... 58
Table 3-6 Mixed NTM and M. tuberculosis clinical isolates ... 60
Table 4-1 MTBDRplus results for crude DNA mixtures of NTM‟s and the MDR-TB isolate. .. 79
Table 4-2 Mixed NTM and M. tuberculosis in clinical isolates... 79
Table 4-3 Speciation and drug resistance genotyping of isolates directly from sputum specimens without culture. ... 80
xii
Table of Contents
Declaration... ii Summary ... iii Opsomming ... v Acknowledgements ... vii List of Figures ... x List of Tables ... xi CHAPTER 1 ... 1 Introduction ... 1 1.1. Background ... 1 1.2. Problem Statement ... 2 1.3. Hypothesis ... 2 1.4. Aims ... 3 Reference List………...4 CHAPTER 2 Literature review ... 5 2.1. Introduction ... 5 2.2. Detection of TB by microscopy ... 62.3. Fluorescent light emitting diode (LED) microscopy... 7
2.4. Drug resistance detection ... 8
2.4.1. Phenotypic methods ... 8
2.4.1.1. Solid Medium Based Methods ... 9
2.4.1.2. Mycobacteriophage based methods for DST ... 11
2.4.2. Liquid Based Methods ... 12
2.4.3. Drug resistance genotyping... 16
2.4.3.1. Preparation of samples prior to PCR amplification ... 16
2.4.3.2. Nucleic acid based methods ... 18
xiii
2.5. Factors affecting drug resistance genotyping ... 22
Reference List ... 29
CHAPTER 3 ... 49
Materials and methods ... 49
3.1. Experimental approach: Project 1- Sputum processing ... 49
3.1.1. Aim ... 49
3.1.2. Bactericide selection ... 49
3.1.3. Bactericides and BCG Pasteur cells ... 50
3.1.3.1. Bactericidal treatment ... 50
3.1.3.2. Killing efficacy after treatment ... 50
3.1.4. Bactericide(s) and M. tuberculosis strains (R439) ... 51
3.1.4.1. Bactericidal treatment ... 51
3.1.4.2. Killing efficacy after treatment ... 51
3.1.5. Bactericide(s) and sputum specimens ... 51
3.1.5.1. Sputum collection ... 51
3.1.5.2. Bactericidal treatment ... 53
3.1.5.3. Killing efficacy after treatment of sputum specimens ... 53
3.1.5.4. Verification of positive cultures ... 53
3.1.5.5. Negative cultures after bactericide treatment ... 54
3.1.5.6. Detection of M. tuberculosis by microscopy ... 54
3.1.5.7. Determination of the DNA integrity by PCR ... 54
3.2. Experimental approach: Project 2- Influence of NTM‟s on drug resistance testing ...57
3.2.1. Aim ... 57
3.2.2. Selection of mycobacterial isolates ... 57
3.2.3. Known mixed NTM and M. tuberculosis in clinical isolates ... 59
3.2.3.1. MTBDRplus assay on sputum specimens... 60
3.3. Experimental approach: Project 3- Hetero-resistance detection with the MTBDRplus assay ...60
xiv 3.4. Experimental Approach: Project 4- Development of a closed line probe assay system 61
3.4.1. Aim ... 61
3.4.2. Development of a device which will prevent contamination during the MTBDRplus assay ... 61
Reference List ... 66
CHAPTER 4 ... 68
Results ... 68
4.1. Bactericidal efficacy ... 68
4.1.1. Bactericidal efficacy on BCG Pasteur cells ... 68
4.1.2. DNA integrity ... 71
4.1.3. Bactericidal efficacy on M. tuberculosis and on sputum specimens ... 72
4.2. Factors affecting the MTBDRplus assay ... 76
4.2.1. Non-tuberculous mycobacteria ... 76
4.2.1.1. MTBDRplus on sputa ... 80
4.2.2. Hetero-resistance... 80
4.2.3. Improving the MTBDRplus assay ... 81
4.2.3.1. Development of a closed line probe assay device ... 82
The device ... 82
Reference List ... 85
CHAPTER 5 ... 86
Discussion and Conclusion ... 86
1
CHAPTER 1
1.
Introduction
1.1. Background
The World Health Organization (WHO), in an initiative for effective Tuberculosis (TB) /Multi-drug resistant TB (MDR-TB) control, developed and adapted the directly observed treatment, short course (DOTS) scheme (8). The main priority of the DOTS was to prevent the emergence of drug-resistant TB through high cure rates of drug susceptible TB. However, as this strategy did not aim at accelerating early detection and subsequent treatment of drug resistant patients, spread and resurgence of drug resistant TB occurred. In recognition of the deadly threat of virtually untreatable TB, the DOTS is now enforced by the WHO to implement rapid drug resistant TB testing by culture coupled with commercially available line probe systems in countries with a high TB burden (12).
In 2006, after 12 years of the DOTS implementation, South Africa‟s (SA) cure rate for new smear confirmed cases, re-treatment cases was 74% and 67% respectively, far below the WHO target of 85% (10). With over 14,000 cases of MDR-TB estimated in 2007 to occur annually, South Africa ranks among the top ten countries in the world for drug resistant TB (11).
Mathematical modelling demonstrates that continuance with the current TB control strategies will exacerbate transmission of TB and drug resistance (7) and that implementation of culture and drug susceptibility testing (DST) on 37% of new TB cases and 85% of retreatment cases will save approximately 50,000 lives, preventing nearly 8000 (14%) MDR-TB cases. However, the model also estimates that there would be no impact on incidence of extensive drug resistant (XDR)-TB (3). A subsequent study also suggests that case detection targets above 70% must be pursued if eradication of TB is to be attained. This may be achieved by increasing diagnosis of TB through active rather than passive case finding and utilizing rapid and highly sensitive techniques such as molecular line probe assays (2).
2 The WHO now recognizes that among response priorities, rapid detection of anti-TB drug resistance, through integration of molecular assays into routine laboratory logarithms as well as monitoring of drug resistance especially MDR/XDR-TB in new patients and its transmission is of paramount importance to curb the disease (11).
1.2. Problem Statement
In 2008 the WHO issued a call for accelerated use of a commercially available PCR based MTBDRplus test for detection of isoniazid and/or rifampicin resistant M. tuberculosis from smear positive sputum cultures (9). Two tests, the Inno-lipa and the MTBDRplus assay were recommended. Evaluation of the genotype MTBDRplus test showed a high correlation with routine culture based DST (1). The MTBDRplus test is now widely implemented in SA for detection of drug resistant TB. However, successful implementation of this methodology requires a change in the way in which M. tuberculosis is inactivated in sputum samples. Prior to microscopy, sputum specimens are currently treated with sodium hypochlorite (NaOCl) to increase the sensitivity of detection of M. tuberculosis by microscopy (4,6). This pre-treatment step also sterilizes the specimen and thereby making it safer for laboratory technicians. However, NaOCl compromises mycobacterial deoxyribonucleic acid (DNA) integrity thereby inhibiting subsequent PCR amplification. This suggests that polymerase chain reaction (PCR) based MTBDRplus test cannot be performed directly on NaOCl inactivated sputum specimens. Thus an alternative method which will allow sterilization of sputum specimen without compromising mycobacterium cell wall integrity for microscopic detection and genomic DNA such that PCR-based drug resistance genotyping can be performed directly on sputum specimens will enable early detection of TB and drug resistant TB. Furthermore, there is a close association between TB and HIV and many immune compromised patients (such as HIV sero-positive patients) can be infected with non-tuberculous mycobacterium (NTM) (5). Due to the growing HIV epidemic, it is not known whether the presence of NTM‟s in sputum specimens influence the detection of drug resistant TB by the MTBDRplus test.
1.3. Hypothesis
In this study we hypothesize that an alternative method which will provide a safe sterile concentrated specimen without compromising mycobacterial stainability, genomic DNA
3 integrity and PCR amplification for genotypic drug DST would significantly improve diagnosis of TB and drug resistant TB. We also hypothesize that a single tube enclosed line probe system will prevent contamination and that the MTBDRplus assay will not be able to detect underlying drug resistant TB in sputum samples where NTM bacterial load is high.
1.4. Aims
To improve the MTBDRplus assay for genotypic drug resistance testing of M. tuberculosis and evaluate the influence of NTM‟s and hetero-resistance on the performance of this test.
Specific aims
To develop methods:
o which will not compromise the ability of inactivated mycobacterium to stain with auramine-O and genomic DNA integrity for PCR amplification.
To determine the influence of the presence of NTM‟s on the detection of drug resistant TB.
To determine the limit of detection of drug resistant M. tuberculosis in hetero-resistant isolates.
To develop a single tube closed line probe system which will prevent contamination
These aims will be discussed separately as different projects: 1.1 Sputum processing
1.2 Influence of NTM‟s on drug resistance testing 1.3 Hetero-resistance
4 Reference List
1. Barnard, M., H. Albert, Coetzee G, O'Brien R, and Bosman M.E. 2008. Rapid Molecular Screening for Multidrug-Resistant Tuberculosis in a High-Volume Public Health Laboratory in South Africa. Am.J.Respir.Crit.Care.Med. 177:787-792.
2. Dowdy, D. W. and R. E. Chaisson. 2009. The persistence of tuberculosis in the age of DOTS: reassessing the effect of case detection. Bull.World.Health.Organ. 87:296-304. 3. Dowdy, D. W., R. E. Chaisson, G. Maartens, E. L. Corbett, and S. E. Dorman. 2008.
Impact of enhanced tuberculosis diagnosis in South Africa: a mathematical model of expanded culture and drug susceptibility testing. Proc.Natl.Acad.Sci.U.S.A. 105:11293-11298.
4. Habeenzu, C., D. Lubasi, and A. F. Fleming. 1998. Improved sensitivity of direct microscopy for detection of acid-fast bacilli in sputum in developing countries. Trans.R.Soc.Trop.Med.Hyg. 92:415-416.
5. Miguez-Burbano, M. J., M. Flores, D. Ashkin, A. Rodriguez, Granada A.M., N. Quintero, and P. A. Noaris Quintero. 2006. Non-tuberculous mycobacteria disease as a cause of hospitalization in HIV-infected subjects. Int.J.Infect.Dis. 10:47-55.
6. Saxena, S., M. Mathur, and V. K. Talwar. 2001. Detection of tubercle bacilli in sputum: application of sodium hypochlorite concentration method. J.Commun.Dis. 33:241-244. 7. Uys, P. W., R. Warren, P. D. van Helden, M. Murray, and T. C. Victor. 2009. Potential
of Rapid Diagnosis for Controlling Drug-Susceptible and Drug-Resistant Tuberculosis in Communities Where Mycobacterium tuberculosis Infections Are Highly Prevalent. J.Clin.Microbiol. 47:1484-1490.
8. WHO. 2005. WHO Global tuberculosis control: Surveillance, Planning, Financing.
9. WHO. 2008. Molecular line probe assays for rapid screening of patients at risk of multi-drug resistant tuberculosis (MDR-TB). Policy statement. 27-6.
10. WHO. 2009. Global tuberculosis control, Epidemiology, Strategy and Financing. WHO/HTM/TB/2009.411.
11. WHO. 2010. Multidrug and extensively drug-resistant TB (M/XDR-TB): 2010 global report on surveillance and response. WHO/HTM/TB/2010.3.
12. WHO/IUATLD. 2008. Global Project on Anti-Tuberculosis Drug Resistance Surveillance: Anti-tuberculosis drug resistance in the world, Report No.4, Annex 9. WHO/HTM/TB/2008.394.
5
CHAPTER 2
2.
Literature review
Current methods for drug resistance testing of Mycobacterium tuberculosis
2.1. Introduction
Mycobacterium tuberculosis, the bacilli that causes Tuberculosis (TB) remains one of the three deadliest infectious pathogens worldwide despite rigorous attempts to control the epidemic by National TB Control programs under the guidance of World Health Organization (WHO) (241). Surveillance data suggest that interplay of several dynamics work in synergy to aggravate acquisition and transmission within communities fuelling the TB epidemic. These include; 1] emergence of multi-drug resistance (MDR) and extensive drug resistance (XDR) TB strains, 2] Human Immunodeficiency Virus (HIV) (3,18,44,68), 3] lack of systemic monitoring of TB cases, 4] delay in diagnosis of new cases and defaulters and 5] inappropriate therapy due to poor treatment adherence or inadequate drug therapy (9,18,62). In 2006, the TB incidence rate in South Africa (SA) was estimated by the WHO to exceed 900/100 000 population per year (238) with more than 6000 new MDR-TB cases detected each year. In 2008, this number escalated to about 13000 MDR-TB cases per year (240,241). Approximately 9.6% MDR-TB, 10.5% XDR-TB cases with 14.2% fluoroquinolone resistance in SA was reported by the WHO in the fourth world drug resistance surveillance (240).
Drug resistance in SA was extensively highlighted by the KwaZulu Natal province outbreak where 72 patients were diagnosed with MDR-TB, of which 53 had XDR-TB (68). The causal strain, F15/LAM4/KZN was found to have been responsible for cases of MDR-TB since 1994 and XDR-TB from 2001 (170). Subsequently it has been shown that other drug resistant TB strains are widespread in the country (144). Outbreaks of MDR-TB have been described in the Eastern Cape [Atypical Beijing] (206), the gold mines [LAM4] (33) and in the Western Cape [Beijing/W-like (95)], [Low Copy Clade (LCC) (228)], [F11] and [F28] (207). The MDR-TB outbreaks in Western Cape are driven by specific strain lineages, the Beijing R220 cluster and the LCC DRF150 strain (207). These strains are highly transmissible irrespective of the presence of characteristic drug resistance causing mutations (119). It is estimated that the incidence of
6 drug resistant TB cases in the Western Cape will double by 2.4 every 8 years. The main driving force of this phenomenon is the increase in cases caused by the Beijing strain R220 (228).
Prior to the 1940‟s, TB drug resistance was thought to emerge only as a result of treatment relapse or inadequate drug therapy, that reactivation of latent infection explained development of disease rather than disease transmission (224). Since the 1950‟s, it has also been thought that drug resistance strains are less virulent and less transmissible, however recent reports provide evidence that drug resistant strains can develop compensatory mutations which restore fitness (33). Certain drug-resistance mutations have also been shown to incur a very low or no fitness cost (23,24). This suggests that outbreak strains have unique properties which aid in increased transmissibility and drug tolerance, however the influence of HIV cannot be excluded (119). Emerging evidence also suggests that the exponential increase over time of the Beijing lineage may be a reflection of enhanced pathogenicity rather than transmissibility (95).
Members of the same bacterial population can independently acquire distinct drug resistance mutations, suggesting that multiple strains can co-exist within a patient with different susceptibility and fitness profiles (209). This phenomenon is defined as hetero-resistance. Hetero-resistance further complicates diagnostic or treatment outcomes as the one strain can become predominant masking the presence and therefore detection of the other strain (177,235). It is therefore important to establish techniques to rapidly and accurately detect drug susceptible and drug resistant TB.
2.2. Detection of TB by microscopy
Conventional light microscopy remains the primary method for diagnosing pulmonary TB in developing countries regardless of its various limitations (130). These include factors such as 1] its value only in areas with high TB incidence or prevalence where patients are diagnosed with high bacterial loads in sputum (5,000-10,000 bacilli.mL-1) (37), 2] its relatively low sensitivity (20-80%) in extra-pulmonary TB and TB/HIV co-infected patients when compared to traditional culture methods (77,102) and 3] the lack of specificity for M. tuberculosis. The problem of the low specificity of smear microscopy has only recently received attention as a result of frequent detection of opportunistic non-tuberculosis mycobacteria (NTM‟s) in HIV co-infected
7 individuals (54,139,167,231,233). Detection and diagnosis of NTM‟s has become particularly important as clinical manifestations of some NTM disease can be indistinguishable from those of TB (46,47). Cases where NTM disease/isolation was misdiagnosed as TB or co-infection with NTM was missed (154) and due to inadequate treatment administration classification of cases as MDR-TB has been documented in resource-limited settings (45,48,101,214). Another major disadvantage of smear microscopy is its limited value for diagnosis in paediatric cases as it is often difficult to get a good quality sputum in sufficient quantities in children (99,152,201).
2.3. Fluorescent light emitting diode (LED) microscopy
Fluorescence microscopy has not gained access to routine diagnostic laboratories in most developing countries despite its higher sensitivity (~10% higher) compared to smear microscopy (37). The cost for routine maintenance as well as running costs (electricity and requirement of specialized mercury vapour light sources) makes fluorescence microscopy less favourable than smear microscopy (140). Several light emitting diode microscopes (LED) which also use fluorescence have been developed (2,140) and recommended to replace fluorescence microscopes and to serve as an alternative to smear microscopy. These microscopes are robust, inexpensive, do not consume too much electricity (124) are portable and can run on batteries (107,140) compared to mercury vapour lights used on fluorescent microscopes. LED microscopes can also be used to view slides at higher magnification (2) and have been shown to reduce time of AFB detection when compared to fluorescence microscopy and are more sensitive than smear microscopy (114,220). Studies evaluating the impact of the implementation of LED microscopes on diagnosis and treatment at point of care sites, as well as combining LED microscopy with novel approaches for early TB and drug resistance case detection are however still required.
Decontamination methods
Strategies to enhance sensitivity of smear microscopy have been assessed. These involve thinning, decontamination and centrifugation of specimens to concentrate bacilli. Numerous decontaminants and digestive reagents such as cetylpyridinium bromide (CPB), cetylpyridium chloride (CDC) have been suggested (239). However N-acetyl-L-cysteine sodium hydroxide
8 (NALC/NaOH) is the most frequently used decontaminant for respiratory specimen prior to microscopy and mycobacterial cultivation (26,239).
Sterilization of respiratory specimens with sodium hypochlorite (NaOCl) prior to microscopy has also been suggested to be an alternative method to concentrate and provide a relatively safe specimen which is non-infectious to laboratory technicians (80). No significant increase in the sensitivity of microscopy subsequent to decontamination with NALC-NaOH and or NaOCl treatment followed by concentration with centrifugation was indicated (10,36,37). High sensitivities were however reported in clinical specimens from a TB/HIV endemic population on specimens processed with NaOH/NaOCl. NaOCl has gained favour in routine microscopic diagnosis in some laboratories as a useful sedimentation agent (80,173).
2.4. Drug resistance detection
Numerous phenotypic and genotypic methods have been described for DST and only a few are frequently used. Although genotypic methods offer several advantages over culture techniques such as reliability, reproducibility, and a short turnaround time (TAT) for results which may potentially help improve patient management (18), DST by culture remains the mainstay to detect drug resistance. Genotypic methods have become particularly important as they enable determination of the specific gene(s) and mutation(s) causing resistance (86) and coupled with phenotypic methods allow determination of resistance (59,204). This chapter will review the currently available phenotypic and genotypic methods for detection of drug resistance.
2.4.1. Phenotypic methods
One of the most important characteristics of culture based methods is that they take advantage of the critical drug concentration which discriminates drug resistant and susceptible strains. The critical concentration is defined as the drug concentration which completely inhibits growth of actively dividing drug susceptible mycobacteria (34). Table 2.1 summarizes the critical concentrations of drugs in different media which form part of treatment regimens frequently used for the treatment of drug susceptible and resistant TB (239).
9 2.4.1.1. Solid Medium Based Methods
Four egg or agar based solid culture methods are proposed by the WHO for DST: these are the proportion, resistance ratio, absolute concentration and the micro-dilution method (237). With the proportion method, the minimum inhibitory concentration (MIC) is defined as the drug concentration where <20 colonies (equivalent to 1% critical proportion) form while confluent growth is observed in the drug free media (34). The MIC in the agar dilution method on Middlebrook 7H10 medium is defined as the drug concentration on which fewer colonies were found when compared with the 10-4 quadrant together with a confluent growth on the 10-2 drug-free quadrants. The laboratory strain, H37Rv is included as an internal control to adjust for batch to batch variation of medium on both methods. The absolute concentration method is similar to the resistance ratio method; serial dilutions of carefully controlled inoculums containing 2 x 103
to 1 x104 CFU of mycobacteria are made (34). These dilutions are then inoculated in media with and without the drug. Resistance is defined as growth that is greater than a certain number of CFU‟s (usually 20) at a particular drug concentration. In these methods, for a resistant isolate the calculated proportion is higher and for a susceptible strain the calculated proportion is lower than the critical proportion (34,237). The proportion method on Lowenstein-Jensen (LJ) is used worldwide as a reference method as DST critical concentrations for both 1st and 2nd line drugs
have been well validated on this media (Table 2.1) (239). A disadvantage of these traditional culture methods is that they take about 4-6 weeks before resistance or susceptibility is confirmed and have limited sensitivity, despite evidence of high specificity (138). Inaccurate or false negative phenotypic results which may result in inappropriate treatment of patients leading to TB related mortalities (176) and the current drug resistant epidemic have also been reported (69). In-house solid medium based culture techniques have been proposed in an attempt to reduce the time for DST results. These include the E-test, nitrate reductase assay (NRA), thin layer agar (TLA) assay as well as mycobacteriophage based assays.
The E-test
The E-test (AB Biodisk, Solna, Sweden) is a commercial system which detects drug resistance on a plastic strip impregnated with a gradient of the antibiotic. This antibiotic gradient also allows determination of the MIC on the surface of 7H11 agar plates supplemented with OADC. The MIC is defined as the drug concentration at which growth of actively metabolizing
10 mycobacteria is inhibited. An inoculum equivalent to 3.0 McFarland is recommended with the results available within 5-10 days. Evaluation studies show correlation when the method is tested against the LJ proportion method for the four 1st line drugs (3,64,65,153). A recent study in
Uganda for rifampicin (RIF) resistance showed complete agreement between the E-test and the direct BACTEC method when compared to the indirect (DST subsequent to mycobacterial culture) BACTEC 460 assay (162). The E-test system is however criticized for the substantial number of false negatives and positives (Table 2.2) (64,76) although two other subsequent studies reported high specificities and sensitivities (3,162). Further evaluation studies are needed for this test.
The nitrate reductase assay
The nitrate reductase assay (NRA), is based on the ability of viable M. tuberculosis to reduce nitrate to nitrite in standard LJ or modified Middlebrook 7H9 medium incorporated with 100 mg/L of potassium nitrate (KNO3) (155). Reduction to nitrite causes a reddish/violet colour
change on the surface of slants on addition of the Griess reagent (11,171). Clinical specimens or cultured isolates are inoculated in the presence and absence of the drug. Adaptation of NRA method for DST directly on clinical specimens has the advantage of reducing the time for availability of results in comparison to the indirect method and has been shown to yield reliable results with high sensitivities for Isoniazid (INH) and Rifampicin (RIF) (230). Pyrazinamide (PZA) resistance detection by the NRA assay with the drug‟s analogue nicotinamide which also possesses an anti-tuberculocidal activity has also been described (126). Both PZA and nicotinamide are pro-drugs that are catalyzed by the enzyme pyrazinamidase/nicotinamidase to pyrazinoic acid (POA) and nicotinic acid (246,247). Nicotinamide unlike PZA which requires an acidic pH works well at a neutral pH and a critical concentration of 250 mg/L produced comparative results to the BACTEC 460 system (126) (Table 2.2). Among other limitations, DST for streptomycin (SM) and ethambutol (EMB) has however been reported to be less reliable with low sensitivities and specificities (Table 2.2). The low sensitivity and specificity to these two drugs may be explained by the presence of hetero-resistant isolates (213) or faster deterioration of the drugs in the media (143).
11 Thin layer agar assay
The thin layer agar (TLA) assay also known as the microcolony method is based on inoculation of an M. tuberculosis suspension on either Middlebrook 7H10 or 7H11 TLA and visualization of microcolonies under a microscope (10X magnification) (192). The TLA assay has been evaluated for the four 1st line drugs (INH, RIF, EMB and SM) and for ofloxacin (OFL) and
kanamycin (KAN) with reliable results (127,180,192) (Table 2.2). The assay is inexpensive, simple to perform and the results are available faster in comparison to traditional solid medium based assays but slower than with the BACTEC systems. Rigorous safety conditions are however necessary (180) (Table 2.2). MDR-TB can be detected within 13 days on smear positive sputum and within 38 days in smear negative sputa (127). Lower contamination rates in comparison to the BACTEC MGIT 960 system, proportion method and the microscopic-observation drug-susceptibility (MODS) assay were reported (Table 2.2).
2.4.1.2. Mycobacteriophage based methods for DST
Two rapid Mycobacteriophage based assays on solid media for M. tuberculosis DST were proposed. The one phage assay is based on Mycobacteriophage replication within M. tuberculosis; replication of phage is then determined by counting viral particles on fast-growing M. smegmatis after overnight incubation. Growth is indicated by clear plaques which form a turbid lawn on M. smegmatis (244). The luciferase reporter format however relies on production of light by recombinant phages on infection into mycobacteria. The recombinant phage contains a luciferase gene which once within viable mycobacteria causes oxidation of luciferin to oxyluciferin and subsequent light emission. Activity of the gene and therefore presence of viable mycobacteria is determined by luminescence in the presence or absence of antibiotics (90). Both formats are based on the replication of phage such as D29, TM4, L5 and Chel2 (16,108,166,187) inside viable mycobacteria.
The FastPlaque response assay also known as the FastPlaque TBTM-MDRi (Biotech Laboratories ltd, Ipswich, UK) is a commercially available Mycobacteriophage replication assay and can be used to detect RIF resistance directly from clinical specimens (7). Decontaminated clinical specimens are inoculated in the presence or absence of RIF then in a suspension of phages. After sufficient time has elapsed to allow the mycobacteria to take up the phage, the
12 remaining extracellular phages not infecting the M. tuberculosis are removed with a virucidal solution. The replicating phages are visualized within 2 days as plaques or lysis when plated in non-pathogenic mycobacterium such as M. smegmatis. Presence of viable M. tuberculosis in the presence of RIF is interpreted as drug resistance to this drug. A major limitation of phage based assays include the generation of inconsistent results in numerous studies (6,7,30,42,66,145,198). The Luciferase reporter phage assay is a high throughput screening micro-plate bacteriophage assay (66). The assay was shown to be highly sensitive and specific for detection of RIF resistance even in an isolate which did not possess mutations in the 81bp region of the rpoB gene suggesting that it could be more sensitive than methods which are limited to the analysis of the hot spot region of the rpoB gene (16,17).
Among limitations of phage based assays, the lack of specificity of Mycobacteriophage for M. tuberculosis, lack of evaluation studies for drug resistance other than RIF, and its inconvenience in testing a large number of isolates are central (Table 2.2). Phage based assays have also been shown to have a TAT and cost comparative to that most molecular based techniques.
Although the E-test, NRA, TLA and Mycobacteriophage based methods are cheaper, they often are not as simple to perform and require high standards of biosafety and quality control (Table 2.2).
2.4.2. Liquid Based Methods
Numerous liquid culture-based DST methods have been introduced in the past years. Many of these assays are commercially available and rely on the principle that M. tuberculosis grows faster in liquid than on solid media. Some of the techniques are manual and require interpretation of results by eye while others are automated.
BACTEC systems
BACTEC system (Becton Dickinson, Franklin Lake, N.J.), the first manual liquid based culture method introduced in the 1970‟s, utilizes modified 7H9 or 7H12 Middlebrook broth containing a
13 acid releasing radioactive 14CO
2. The amount of radioactive 14CO2 released is translated into a
numerical value designated as the growth index (GI), a GI value higher than 10 is considered as positive (196). The first evaluation on the rapid radiometric DST of M. tuberculosis was conducted in 1981 (196,200). A subsequent study reported overall agreement of radiometric results with those obtained by the proportion method with specificities, sensitivities higher than those stated in an earlier study. The BACTEC 460 system has since been the leading rapid culture system for M. tuberculosis for the past two decades and provides DST within 4-12 days (69,186,189,200,218). Critical concentrations for both 1st and 2nd line drugs have been validated on this system (Table 2.1). The BACTEC 460 system is also the reference method for DST for PZA resistance as a BACTEC 460 vial is available which provides the necessary acidic medium (pH 5.9) for the activity of the enzyme (92).
The BacT/Alert 3D system
Another addition to the BACTEC system is the BacT/Alert 3D system formerly known as the MB/BacT system (bioMireux, Durham, N.C). It is an automated commercial liquid culture-based system that measures microbial growth every 10 minutes (188). The system uses a modified 7H9 Middlebrook broth containing a pH indicator mixture and incorporates a colometric sensor at the bottom of the vial which measures changes in CO2 production by the metabolizing mycobacteria.
Elevated CO2 concentrations lower the pH of the medium which in turn produces a colour
change in the sensor which is detected by the reflectometric unit of the instrument (236). Change in colour from green to yellow indicates a positive reaction and each vial is continuously monitored inside the apparatus (12,169). Evaluation studies show concordance with reference methods with comparable sensitivities and specificities (Table 2.2).
The BACTEC Mycobacteria Growth Indicator Tube (MGIT) 960 system
The BACTEC MGIT 960 (Becton Dickinson Microbiology Systems, Sparks, MD) is an automated version of the MGIT system. The systems are based on an oxygen-quenching fluorescence sensor embedded at the bottom of the tube containing enriched Middlebrook 7H9 broth. On the BACTEC MGIT 960 system incubation and reading of fluorescence emitted occurs continuously inside the machine using a predefined algorithm to interpret the fluorescent signal and gives results as either negative or positive (190). Recently included in the BACTEC MGIT
14 960 system is the MGIT 960 PZA kit containing tubes with modified 7H9 broth which enhances growth of mycobacteria at pH 5.9 which enables detection of PZA resistance (92). In DST, the BACTEC MGIT 960 system interprets the results as susceptible or resistant to the antibiotic under investigation with results available in 8 days. Although this system is used more regularly than the BACTEC TB-460 it poses more danger to laboratory personnel and requires special disposal systems. The BACTEC MGIT 960 system however is reported to be more prone to contamination than in the BACTEC 460 system (Table 2.2) (242). The performance of the BACTEC MGIT 960 system has also not been shown to be superior to that of the BACTEC 460 system (242).
Recently, the focus has shifted to rapid and affordable direct tests in which clinical specimens are directly inoculated in drug-free and drug containing medium (41,74) or amplified for detection of drug resistant-TB (79,104). These assays include the MODS assay, microplate alamar blue assay (MABA), tetrazolium microplate assay (TEMA), resazurin microplate assay (REMA) (53,147,164) and the ESP II system (110).
Microscopic Observation Drug Susceptibility assay (MODS)
MODS is an „in-house‟ assay which enables DST by detection of early growth of M tuberculosis as „strings and tangles‟ of bacterial cells in 7H9 Middlebrook medium (38). The inoculation is performed on 24-well plates with or without antibiotics at 37°C (149). The assay has numerous features which make it suitable for use in resource poor settings; however several aspects also limit its application in these settings (Table 2.2). MODS is not superior to any liquid culture based assay and is also prone to contamination in comparison to other liquid culture systems which necessitates repeated or rigorous decontamination of specimens (57,60,70,150).
The versa TREK system
Previously known as the ESP culture system ΙΙ (Trek Diagnostic systems, West Lake, OH) this non-radiometric automated method can be used for detection and DST of M. tuberculosis. This system along with the BACTEC 460 was the first two broth systems accepted by the Food and Drug Administration for DST. Using an enriched 7H9 broth, the assay detects mycobacterial growth by measuring gas pressure changes inside culture vials caused by mycobacterial
15 metabolism. The ESP II system can rapidly detect drug resistance to INH and RIF but discrepancies between the system and BACTEC 460 for STR and EMB were observed (20,182). PZA resistance with the BACTEC 460 system as the reference was in concordance (110). No evaluation studies have yet been published for performance of this assay directly on clinical specimens.
Colometric microplate liquid medium based assays
Colometric in-house assays were developed with the aim to overcome the high cost of commercially available techniques (164). These methods rely on colour changes which are observed by eye resulting from oxidation-reduction (redox) reactions. The redox reactions occur between an indicator and O2, CO2, NO3 or drug metabolism in culture medium when M.
tuberculosis is grown in the presence or absence of an antibiotic. These assays include, methods such as the MABA, TEMA and the REMA (128,158,164,175). The alamar blue, 4, 5-dimethlythiazol-2yl)-2, 5-diphenyl tetrazolium bromide (MTT) and resazurin are used as indicator dyes for MABA, TEMA and REMA assays respectively (116,142,147).
TEMA and MABA have been shown to be highly sensitive and specific for detection of 1st line
drugs (Table 2.2) (115,147). Discordances for SM and especially EMB have been reported. These discrepancies have been attributed to differences in the media of the respective assays or degree of degradation of the respective drugs in the media (136). DST for 1st line drugs as well
as PZA resistance with nicotinamide has also been determined on the REMA assay (129) (Table 2.2). Nicotinamide at the critical concentration of 250 mg/L on REMA assay produced comparative results to the BACTEC 460 system (129). As colour change can be detected visually, no additional devices are necessary and growth can be detected earlier before colonies become visible (147). REMA is more cost effective compared to MABA and TEMA as resazurin (the main component of alamar blue) is less expensive than alamar blue or MTT (142,147). Although these colometric assays are cost effective, rapid and also allow MIC determination, low sensitivity and specificity for EMB, STR and capreomycin (CAP) is documented (Table 2.2). The need for rigorous safety conditions for lab personnel may also limit their use in resource limited countries.
16 2.4.3. Drug resistance genotyping
Genotypic methods for DST are based on evidence that acquisition of drug resistance mutations in M. tuberculosis occur by chance as a result of imperfect chromosomal replication. Bacteria with drug resistance mutations are positively selected during periods of discontinuous drug therapy (e.g. through non-compliance or inappropriate therapy) (94). Resistance causing mutations occur mainly in regions targeted by antibiotics, in enzymes which activate or deactivate certain essential genes or on the promoter regions upstream of these genes. Resistant bacteria which do not have classical drug resistance mutations suggest unknown mechanism(s) are involved (94,248). This is also supported by the differences observed in MIC‟s of strains harboring the same mutation (87,94,122).
The mutation rate on the M. tuberculosis genome as determined using the Luria-Delbrûk fluctuation analysis demonstrates that in vitro INH-resistance emerges at a rate of 3.5 x 10-6 and
RIF resistance at a rate of 3.1 x 10-8 mutations per cell division (55). DNA sequencing has shown
that 95-98% of RIF-resistant isolates are also resistant to INH, making RIF-resistance a good predictor for MDR-TB (156). More than 90% of M. tuberculosis strains phenotypically resistant to both RIF and INH respectively were also shown to harbor point mutations within the 81bp-hot spot region of the rpoB gene (codon 507-533). As a result of these findings, detection of RIF resistance conferring mutations within this region forms the foundation for the detection of MDR-TB. More recently, due to the resurgence of XDR-TB, evidence of INH (197) and RIF mono-resistance (207), multiplex PCR assays which incorporate classical mutations causing resistance to both 1st and 2nd line TB drugs have been developed (86). Prior to genotypic drug
resistance testing, smear positive sputum specimens are decontaminated followed by mycobacterial cultivation or direct genotyping on sediments. The decontamination step is crucial as it sterilizes the specimen and increases mycobacterial detection.
2.4.3.1. Preparation of samples prior to PCR amplification
The standard protocol for preparation of clinical specimens prior to PCR amplification involves digestion and decontamination of specimens with NALC-NaOH (61). DNA extraction methods by commercial kits, proteinase K phenol-chloroform purification method, carboxypropylbetaine (CB-18) coupled with glass beads homogenation have also been explored (50,100,212,234).
17 Although these techniques provide purified DNA, multiple steps which are time consuming, labour intensive and expensive are involved (212). The boiling method which involves mycobacterial cell lysis for 5 minutes at 100ºC to release crude DNA and reduce PCR inhibitors is a simple, cost effective and rapid method which has been shown to yield sufficient DNA for PCR (212,217).
Bactericidal reagents which sterilize and simultaneously lyse mycobacterial cells have not yet been explored for PCR based diagnostic techniques. There are commercial kits which use bactericidal reagents coupled with PCR based assays, however the bactericides in these kits are not mentioned (15). Limited data is available on the efficiency of bactericidal reagents against mycobacterial species. Literature reviews focus on bactericides against a variety of microbial organisms but not specifically on M. tuberculosis. The most frequently used bactericides are aldehyde based sterilizing reagents such as glutaraldehyde (GTA), NaOCl and peracitic acid (C2H4O3) and are mainly used for sterilization of medical equipments and hospitals (71).
More recently ortho-phthalaldehyde (OPA), an aromatic dialdehyde has been proposed as a possible alternative to GTA for high-level disinfection (211). It was shown that 0.5% w/v OPA can rapidly and efficiently sterilize a range of NTM‟s and more importantly GTA-resistant mycobacterial strains (211). Although GTA is a more effective cross-linking agent and its own uptake may be decreased by virtue of its extensive cross-linking nature to amino acid residues at the bacterial cell surface. GTA thus inactivates cells at a slower rate than OPA at the same concentration (63). It is suggested that OPA may induce inactivation by cross linking the active sites of cysteine to neighbouring lysine residues (199).
NaOCl, with hypochlorous (HOCl) acid as the active moiety, is a powerful bactericidal, sporocidal and fungicidal agent which has been demonstrated to be highly effective against M. tuberculosis (174). It is relied upon for sterilization of mycobacteria by lysing the mycobacterial cell through interactions with amino acid groups and primarily through progressive oxidation of disulphide resulting in consequent degradation of DNA within minutes of contact (10). HOCL is a potent bactericidal agent which is active at even concentrations below 0.1mg/liter (174). Glucoprotamin is non-volatile, water soluble, non-corrosive, non-mutagenic, and easily
18 degradable compound (243). It is bactericidal against mycobacteria, fungi and viruses. It is highly active against glutaraldehyde resistant strains of M. chelonae (137).
Peracitic acid, an oxidizing agent effective at low temperatures, decomposes to non toxic residues and is effective in the presence of organic matter. The extra oxygen atom is highly reactive and interacts with most cellular components and functions to cause cell death. Peracitic acid inactivates many different critical cell systems and this is the key to its broad spectrum antimicrobial activity even against resistant M. terrae and M. avium (205,221).
A challenge with bactericide sterilization prior to PCR amplification may be however its implications on DNA integrity or DNA degradation and possible introduction of sequence changes or false positives on genotypic results (15,199). Determination of the time duration for the respective bactericide to penetrate the mycobacterial cell wall, efficiently sterilize the clinical specimen or culture without causing sequence changes or DNA degradation is critical.
2.4.3.2. Nucleic acid based methods
Mutations in one or several genes with different mutation frequencies are implicated in the acquisition of drug resistance in M. tuberculosis (Table 2.3). Drug susceptible isolates lack these corresponding gene mutations and this forms the basis of drug resistance genotyping. The inability to detect all mutations conferring resistance remains a major challenge to the successful development of highly sensitive genotypic DST methods. The main reason for this is that the mechanisms of drug resistance to some of the drugs are not fully understood (94). Different geographic regions with different prevalence rates of mutations in the drug resistance conferring genes have been documented and this further complicates the development of these techniques (39,98,184). Region specific mutation screening methods are not yet in place and this may lead to misdiagnosis of drug resistance. Many of these methods are also complicated by the need for downstream processing to enable the detection of genotypes within the amplified PCR product. The complexity and multiple steps of these techniques greatly increase the risk of cross-contamination and thereby misdiagnosis.
19 Single Strand Conformational Polymorphism (SSCP), Restriction Fragment Length Polymorphism (RFLP), allelic Amplification refractory mutation system (ARMS)-PCR, heteroduplex analysis (HA), pyrosequencing (84) and probe based hybridization methods such dot blot analysis (227) have been described. However, RFLP-PCR based DST has limited use as not all mutations conferring drug resistance results in gain or loss of restriction enzyme cutting sites (229). ARMS, HA and SSCP-PCR (97) are technically challenging, time consuming with limited sensitivity due to high G-C rich regions in mycobacterial genomes which may influence mobility shifts of DNA sequences (157). Pyrosequencing, an alternative sequencing method to direct DNA sequencing is a rapid high throughput technique that relies on real time detection of pyrophosphates (PPi) (89). In contrast to standard sequencing, pyrosequencing does not make use of fluorochromes, radioactivity or need post-reaction processing of PCR-products. Pyrosequencing approach that combines automated real-time (RT) PCR amplification with pyrosequencing has been devised for detection of drug resistance to INH, RIF and fluoroquinolones (27,75). Although pyrosequencing is high throughput similarly to standard sequencing, it is also cumbersome, costly and technically challenging.
The most promising in-house and commercially available molecular methods developed to detect drug resistant TB will be reviewed in the next few sections.
Direct DNA sequencing
For the last decade, this method has been incorporated into the work flow of a number of clinical mycobacteriology laboratories and is the reference method for detection of drug resistance mutations (85). It involves amplification of the gene of interest or a specific region associated with resistance causing mutations and subsequent sequencing of the amplified product to determine the presence or absence of specific mutations (85). Sequencing is not only costly, with the need for expensive equipment, but also requires expertise (227). Continued use of PCR-DNA sequencing for routine DST is therefore impractical in resource poor settings where cost effective, rapid techniques would be more of value (227).
20 2.4.3.3. Hybridization Based Methods
In these assays, amplified PCR products of genes known to confer drug resistance are reversely hybridized to immobilized nitrocellulose membrane bound allele-specific labelled probes complementary to the wild type or mutant sequence of the gene. Hybridization can be visualized by autoradiography, enhanced chemiluminescence, alkaline phosphatase or other detection systems (111,162). These assays include the commercially available INNO-Lipa RIF-TB test, MTBDR, MTBDRplus and the newly developed MTBDRsl assay.
There are two commercially available assays recommended by the WHO: the Inno-LiPA Rif.TB (Innogenetics, Belgium) and the Genotype MTBDRplus assay (Hain-Lifescience, Nehren, Germany). These assays are based on multiplex PCR amplification of nucleic acid segments and subsequent reverse hybridization of amplicons onto nitrocellulose membrane with immobilized probes specific for M. tuberculosis complex and mutations responsible for drug resistance.
Inno-Lipa Rif.TB
The Inno-LiPA Rif.TB (Innogenetics, Ghent, Belgium) targets the 16S-23S ribosomal RNA (rRNA) spacer region for differentiation of the M. tuberculosis complex and detects mutations within the hot spot region (codons 507-533) of the rpoB gene (91,194). A major limitation of this technique is that it can only detect RIF resistance (91,162,194,203).
Genotype MTBDRplus assay
The MTBDR (Hain Lifescience, Nehren, Germany) assay enables simultaneous detection of the M. tuberculosis complex, the most common mutations conferring RIF resistance and mutations in the katG for INH resistance. The Genotype MTBDRplus, a latter version of the MTBDR includes mutations in the promoter region of the inhA gene for INH resistance, which also confers cross resistance to ethionamide (ETH) (88,191). Both versions accurately identified RIF resistance in 98.7% of the cases, when compared to phenotypic DST (18,81,82). Furthermore, the Genotype MTBDRplus reported higher sensitivity for INH resistance (Table 2.4). The MTBDRplus assay was reported to perform well with significant readability in clinical specimens (160). The assay‟s reliability however can be hampered by presence of rare mutations and hetero-resistance in clinical specimens (160). Despite these limitations when performed
21 directly on clinical specimens, the MTBDRplus assay has been shown to be a reliable, reproducible line probe PCR-based method which could significantly increase early detection of drug resistant TB (18,31,141,160).
MTBDRsl assay
The MTBDRsl assay (Hain Lifescience, Germany), is the latest version of the MTBDRplus assay. It is based on the same principle as the MTBDRplus except that it detects resistance to 2nd line drugs; fluoroquinolones, aminoglycosides (amikacin and kanamycin) and the cyclic peptide; capreomycin and ETH (104). Three recent studies show sensitivities >75% for aminoglycoside and capreomycin resistance detection and even lower sensitivities for EMB resistance (Table 2.4) (28,83,104). This is currently one of the most promising genotypic assays for rapid detection of 2nd line drug resistance.
MLPA assay
The multiplex ligation-dependent probe amplification (MLPA) assay was initially designed to screen copy number changes and CpG methylation changes in human genomic DNA (161,226) and characterize bacterial genomes. Due to its high probe capacity (can incorporate ~45 probes), this assay was modified to enable simultaneous detection of resistance to RIF, INH, and EMB as well as genotype specific mutations in the M. tuberculosis genome for specie identification in a single assay (21). The selected drug resistance markers can detect approximately 70-85% RIF, 65-80% INH and 45-65% EMB resistance (21). M. tuberculosis genotype specific speciation with simultaneous drug resistance genotyping may help detect outbreak strains early and may provide valuable information for treatment and prevention of TB transmission. No evaluation studies in large trials have been performed and this needs attention.
Real- time PCR
RT-PCR methods which make use of Molecular Beacon (MB), TaqMan minor groove binder (MGB) and fluorescence resonance energy transfer (FRET) probes have been described (105,181,232,245). The sensitivity and efficiency of MB and TaqMan probe coupled PCR at detecting MDR-TB in DNA extracts with mixed ratios of mutant and sensitive sequences was determined (245). The detection efficiency was shown to be determined first by the amount of
22 each sequence and then by the ratio of the sequences irrespective of the probe system. The presence of a second allele did not influence specificity of the probe system (245). In this study FRET probes were unable to produce an amplification signal in mixed sequence specimens regardless of the quantity or ratio of the different sequences (245).
Cepheid GeneXpert MTB/RIF assay
The Cepheid GeneXpert MTB/RIF (Sunnyvale, CA) assay a MB RT-PCR based assay has recently been described for detection of RIF resistance (79). This assay is an enclosed, fully automated system which processes, extracts, purifies and amplifies target sequences directly from clinical specimens. The Cepheid GeneXpert MTB/RIF assay makes use of a single use sample-processing cartridge system with the GeneXpert instrument which has an integrated multicolour real-time PCR detection capacity (79,222). It is the first assay which incorporates sample processing, amplification and nucleic acid analysis in one tool (183,222). The assay makes use of 6 MB probes with their respective fluorescent dyes and quenchers for detection of RIF resistance conferring mutations and has been shown to be highly sensitive and specific even on smear negative specimens (26,79). It has also been shown to detect M. tuberculosis H37Rv in sputum specimens with mixed NTM species. The GeneXpert has several demonstrated advantages such as its large dynamic range, high specificity and on site diagnosis of MDR-TB while patients wait (79). The authors showed that the amount of aerosols produced are comparative to those generated during smear microscopy and pose insignificant threat to laboratory personnel (15). The sterilizing reagent was reported not to influence the sensitivity of the assay after 3 hours of exposure, however RIF false positive results were reported (15). Although expensive, the GeneXpert MTB/RIF assay is by far one of the most promising recent genotypic tools introduced for DST. Further evaluation of this assay for simplicity, robustness and accuracy for immediate DST performance in health care facilities to enable appropriate drug administration to TB suspects is required (26,79).
2.5. Factors affecting drug resistance genotyping
Numerous factors such as the DNA concentration, presence of PCR inhibitors, silent or excluded mutations and contaminants may affect the specificity and sensitivity of molecular methods; some of the factors are summarized in Table 2.5. Contaminants can be overcome by rigorous
